Real-time MHE-based nonlinear MPC of a Pendubot system ⋆

This paper addresses the real-time optimal control of a Pendubot using nonlinear model predictive control (NMPC) combined with nonlinear moving horizon estimation (MHE). This fast, under-actuated nonlinear mechatronic system apparently poses a challenging benchmark problem that may benefit from a nonlinear optimization scheme. To overcome the related computational difficulties we make use of the ACADO Code Generation tool allowing to export a highly efficient Gauss-Newton real-time iteration algorithm tailored to the nonlinear optimal control and estimation problem while respecting the imposed constraints. We show experimental results illustrating the overall closed-loop control performance, as well as the advantages of the nonlinear MHE-based NMPC scheme.